With the ever developing of modern network technology and broadband construction, a videoconference system has been extensively applied to areas of work and life. In a television conferencing application, MCUs can be cascaded to one another. In a cascade conference, terminal conference sites under different MCUs can be scheduled in the same conference, that is, the networking in a cascade manner shown in FIG. 1 exists. An MCU2 and an MCU4 are hung under an MCU1. An MCU3 is hung under the MCU2, and an MCU5 is hung under the MCU4. Conference sites from T31 to T3m are the conference sites hung under the MCU3, and conference sites from T51 to T5n are the conference sites hung under the MCU5 (both m and n are integers greater than 1). When the video conferencing system shown in FIG. 1 is used to hold a cascade MCU conference, because many MCU cascade levels (more than two levels of MCUs) exist, media stream delay due to the MCU cascade and loss resulting from multiple encoding and decoding seriously affect video and audio effect. For example, as shown in FIG. 1, the conference site T51 hung under the MCU5 receives a media stream from the T31 under the MCU3. Because multiple forwarding by MCUs is performed during the process (MCU3->MCU2->MCU1->MCU4->MCU5), the transfer path is long. Therefore, when the conference site T51 receives the data from the conference site T31, the video and audio is delayed for a long time. At the same time with audio mixing of all levels of MCUs and multiple times of encoding, video and audio quality is also greatly lowered.
In the conventional art, some solutions as follows are used to improve the lines between cascaded MCUs. For example, improvement of lines, such as the changing of the materials of lines between MCUs is implemented to achieve the objective of alleviating delay of a media stream and a decline in video and audio quality due to transmission. However, the improvement of the transmission lines is of high input costs, and is not easy to be implemented. As for the effect of improvement, delay caused by multiple forwarding by MCUs cannot be handled, and therefore this solution is not effective on solving the overall delay.